Last week, we started learning about generative design and topology optimizations in Autodesk Fusion 360. The assignment was to design a functional chair by assigning basic surfaces and forces. I wanted to generate something pragmatic and, since I was working on a pot already, I decided to generate a potholder with a water tray.

Experimentations:

Printing big scale items in parts and assembling (pot: 48x9x10 cm, tray: 50x11x6 cm)
Using multiple materials (wood/gray/black PLA)
Experimenting with generative design and shape optimization

Supplies

Rhino

Grasshopper

Fusion 360

3D-Printer

Step 1: Pot Design in Rhino

First, I produced a very basic pot: I created an open box in Rhino (48x10x9 cm) and split it into 3 parts. I didn't split the box equally because I wanted to use the full height of Ender 3 Pro (>20cm). Therefore, the parts at the ends are ~20 cm, and the middle piece is ~10 cm in width.

I also created a jagged joinery system to be able to attach parts together. I manually drew the joinery and drainage curves, extruded them, and used MeshBooleanDifference/MeshBooleanIntersection commands in Grasshopper to create the final shape. I left a 0.3 mm gap in between parts, but after the print, I realized it wasn't enough. I had to sand off a thin layer on either side, and it worked (after breaking one brittle connection).

I printed the end pieces in wood color, and middle piece in gray. Pieces are printed vertically to use as little support material as possible. Total print time was around 2 days.

Step 2: Pot Tray in Fusion 360 (Generative Design | Setup)

Originally, I was going to design a simple pot tray in Rhino and Grasshopper, but I wanted to use this opportunity of simulating a structure that generates a mesh with optimum load distributions. This was interesting because I will ultimately use the pot, and wet soil will be quite heavy.

I have done 7 studies with the Generative Design tab in Fusion 360. I don't know if I was making a mistake somewhere but each result took more than 12 hours to finish. In each study, I modified my problem setup (i.e. constraints, loads, geometry to preserve and occlude). In the final study, I used the setup shown in the first figure.

I started with a water tray that will keep drained water from the pot. I needed some parts to elevate the pot and provide support from all ends. I created floating holders and applied the force on these surfaces. I used 200 N force for the center supports, and 150 N for the sides. Then I sketched four-leg seeds that will sit on the surface of the window ledge. Also, I measured and modeled the window corner itself as a base to be able to use the full space available at such a cramped location. All these bodies are marked as preserved (green). In addition, I roughly modeled the window surface, ledge, and the pot itself to use as obstacles (red).

After creating preserve structures, I needed to test if it actually fits on my window ledge. In Cura, I created a new profile for ultra-draft printing, by increasing the layer height to 0.3 mm and increasing the travel speed of my nozzle head (above figure shows the settings). I printed a tiny crosssection with this setting and found out that my measurements were off. I fixed the dimensions and was able to print a good one that will fit.

Step 3: Pot Tray in Fusion 360 (Generative Design | Explore)

You can see interesting looking designs across 7 studies in the second figure. There were 3 main features that I could see from the best results: (1) wave-like support structures on the back of my model, (2) podium-like holders, and (3) toilet bowls (!everywhere!). The last one, in particular, is quite unfortunate because I cannot unsee the similarity of my potholders with toilet bowls!

I ended up exporting Study 9 - Outcome 2 as it contains all the above main features that emerged from my problem set.

Step 4: Pot Tray in Fusion 360 (Model Cleaning & Splitting)

As the last step, I exported the model as a new design. My initial idea was to draw solid geometry based on the high poly mesh, but this was more daunting then I imagined. Working with curves wasn't as straightforward as I imagined it would be. So instead, I reduced the number of face to a manageable range and converted the model to solid geometry. I simplified some parts (mostly contact points) of the model.

Since the model is quite large, I needed to somehow split it into three. I utilized the thickest part of my design in the center (with two bases) and created a sliding latch system by simply sketching the shape and extruding (2nd figure). I left a 0.5 mm gap this time; I hope it will be enough for Ender 3 Pro.

Step 5: Pot Tray Printing

Printing the tray will be more challenging than printing the geometric pot. Points of consideration:

What's the most optimal way of placing these pieces on the bed?
Do I have enough material? (I have 488grms of black filament with the spool -- does anyone know how much the spools weigh?)
Tree support structures (Thanks @kobernik)
Different infill rates for legs and the rest
- In Cura, add a "Support Blocker" cube anywhere in the space, then select the cube, choose "Per Model Settings", select "Modify settings for overlaps". Then just select infill density and mark it higher or lower.
Should I use print acceleration? Should I increase the print speed of supports?

The average print times I have seen so far with all 3 pieces on the bed are about 2.5-3 days. I will try my best to lower this and start printing.

Update #1: I just realized that the tree support structures of different parts on my printer bed can merge together and reduce the amount of support material required.

Update #2: I found out that the empty spool weights about 250grs. It means I will have to use gray filament for printing the tray.

Step 6: Planting

My original intention was to plant a Devils Ivy in here, but the plants I currently have are already root locked and they are bigger than the width of my pot. So I wanted to create a little succulent garden along with various rocks. I put some rocks inside the soil as well to make the whole structure more grounded.